The annotation and function of the Parkinson's and Gaucher disease linked gene GBA1 has been concealed by its protein coding pseudogene GBAP1

The human genome contains numerous duplicated regions, such as parent-pseudogene pairs, causing sequencing reads to align equally well to either gene. The extent to which this ambiguity complicates transcriptomic analyses is currently unknown. This is concerning as many parent genes have been linked to disease, including GBA1, causally linked to both Parkinson’s and Gaucher disease. We find that most of the short sequencing reads that map to GBA1 , also map to its pseudogene, GBAP1 . Using long-read RNA-sequencing in human brain, where all reads mapped uniquely, we demonstrate significant differences in expression compared to short-read data. We identify novel transcripts from both GBA1 and GBAP1 , including protein-coding transcripts that are translated in vitro and detected in proteomic data, but that lack GCase activity. By combining long-read with single-nuclear RNA-sequencing to analyse brain-relevant cell types we demonstrate that transcript expression varies by brain region with cell-type-selectivity. Taken together, these results suggest a non-lysosomal function for both GBA1 and GBAP1 in brain. Finally, we demonstrate that inaccuracies in annotation are widespread among parent genes, with implications for many human diseases. ### Competing Interest Statement S.S., Y.G., J.E., H.S. and C.F.B. are employed by Astex Pharmaceuticals. The other authors declare no competing interests.